English

Cluster-based Input Weight Initialization for Echo State Networks

Machine Learning 2022-05-11 v3 Artificial Intelligence Neural and Evolutionary Computing

Abstract

Echo State Networks (ESNs) are a special type of recurrent neural networks (RNNs), in which the input and recurrent connections are traditionally generated randomly, and only the output weights are trained. Despite the recent success of ESNs in various tasks of audio, image and radar recognition, we postulate that a purely random initialization is not the ideal way of initializing ESNs. The aim of this work is to propose an unsupervised initialization of the input connections using the KK-Means algorithm on the training data. We show that for a large variety of datasets this initialization performs equivalently or superior than a randomly initialized ESN whilst needing significantly less reservoir neurons. Furthermore, we discuss that this approach provides the opportunity to estimate a suitable size of the reservoir based on prior knowledge about the data.

Keywords

Cite

@article{arxiv.2103.04710,
  title  = {Cluster-based Input Weight Initialization for Echo State Networks},
  author = {Peter Steiner and Azarakhsh Jalalvand and Peter Birkholz},
  journal= {arXiv preprint arXiv:2103.04710},
  year   = {2022}
}

Comments

Accepted for publication in IEEE Transactions on Neural Network and Learning System (TNNLS), 2022